single point z-correction after 3-point auto-leveling
Photo Credit: Courtesy of 3DStockPhoto ( image)

There is no "standard" g-code, especially for the auxiliary functions of calibration, and even more so for the calibration of delta-class printers. I find that the Reprap Wiki includes a fairly comprehensive list: http://reprap.org/wiki/RepRap_Firmware_G-Codes and http://reprap.org/wiki/G-code. These lists address the RepRap firmware, although they include some g-code functions which are specific to other types of applications.

If the Monoprice Delta firmware is based on the RepRap firmware, the lists might be helpful, although detailed documention from Monoprice would be the best reference.

Assuming there is no available or known g-code to adjust this, you work from the hypothesis that your three switches are not responding identically. If possible, use a micrometer to measure the difference between the "idle" position of the bed and the point where the switch actuates. It could be that the front switch requires more travel before actuation, which would then cause the leveling calculation to have a lower point for the front switch. In the alternative to measuring, it may be possible to exchange the front switch with a rear/side switch and see if the problem moves.

Delta machines are more difficult to calibrate because everything interacts. The three actuator delta machine is a simplification of a more complex 6-axis Stewart platform. As I see it, a delta machine is a 6-actuator machine in which pairs of actuators operate "identically", except that minor variations cause unexpected positioning errors. Sometimes these are modeled as a bowl shape with positive or negative curvature, although measurements I've made on my home-brew delta show an error that looks more like a carpet in a high wind. Some factors improve the design, such as longer actuator arm pairs and smaller build diameter, which may help you with the Monoprice.